Kernelized FIR Model With Sparse Approximation for Scalable Soft-Sensing of Wheel Center Load in Vehicles

View at publisher Summarise with VIRAT
Authors: Xu Ge, Mian Li, Yu Qiu, Mingming Zhang
Cited by: 0
References: 27

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Wheel Center Load (WCL) is crucial for assessing the system durability in vehicle health monitoring (VHM). Advanced WCL estimation methods are expected for the substitution of WCL sensors due to high cost. A traditional way is building a data-driven model based soft-sensor system for vehicles. However, linear models cannot estimate high-quality WCL data due to the nonlinearity in working conditions and vehicle systems, while complex nonlinear models require very long training time and large computational resources. This paper aims to design and implement a nonlinear soft-sensor system to estimate WCL in real time with limited computing resources. In this paper, we propose a Kernelized FIR (KFIR) model to estimate the Wheel Center Loads, which demonstrates superior predictive capabilities compared to the conventional FIR models. Kerneling the linear FIR model allows it to efficiently handle nonlinear relationship by mapping input features into higher-dimensional spaces. Then, a greedy algorithm-based sparse approximation method is proposed and developed to expedite the training and predicting processes. The scalable soft-sensor system is built on a prototype vehicle and implemented on many testing vehicles in a large scale. Experimental results indicate that the proposed method achieves a high estimation accuracy with limited computing resources on multiple testing vehicles.

Fields of Study & Subjects

Vehicle Dynamics and Control Systems Fault Detection and Control Systems Transport Systems and Technology

Publication details

Journal / source
IEEE Transactions on Vehicular Technology
Volume
75
Issue
8
Pages
16204–16219
Publication date
31 March 2026
Language
EN